In 1932, Otto Wacker got 19 months for selling 33 fake Van Goghs good enough to fool the experts. The leading Van Gogh scholars signed certificates for them. In court, it was the painter's nephew who undid him: the family's records had no trace of the Russian collector Wacker said they came from.
So why can a real Van Gogh sell for $117 million while a perfect copy is worth next to nothing? Not the technique. A good forger has that. It's who made it, and the story behind it. And Van Gogh sold almost nothing while he was alive. His sister-in-law, Jo van Gogh-Bonger, spent decades publishing his letters and building his audience. The story needed someone to carry it.
Last Tuesday, Michael Smith got 18 months for the same trick with music. AI made him hundreds of thousands of songs, up to 10,000 fake accounts played them billions of times, and he collected $8 million. That's more than most real musicians earn in a lifetime.
When making gets this easy, the business model cracks. More than half the songs uploaded to Deezer every day are now made by AI. Streaming pays from one shared pot, and the three major labels take about two thirds of it. Real artists lost out, of course. But most of that money was the majors', and their whole model depends on making music being hard.
The same week, a solo developer used AI to rebuild Photoshop, Premiere, Lightroom and four more Adobe apps, and gave them away for free. Part of the push was Adobe's cancellation fees. They're rough, but he reckons parity is months away, not years. "Software is over," he said.
Maybe. Either way, the technique is free now and the tools are following. What can't be copied is your audience.
The creator economy is the moat!
🎵 Milli Vanilli — Girl You Know It's True The original fake act. Millions of records, then the only Grammy ever taken back. Full playlist and archive at niccjohnson.com/drop.